36 replies to this topic

[ccaissie]

So previous thread on the Unitron 4" f/15 revealed strong undercorrection...Star test, Ronchi, and now DPAC. Spacer change=nada.

 

Foucault shows the inverse of the hyperbolic/strongly overcorrected pattern that I have seen in mirrormaking.  Have never seen such an oblate figure.

 

So the lens goes back to the owner...but gee, it SHOULD be better.

 

So R4 is coated...don't want to touch that.  What's the easiest surface to alter, and what's the direction to go in?  I assume I'd use Foucault to get a reading of R3, and correct that surface. 

 

In OSLO, if I parabolize or overcorrect the concave R3, that overcorrects the objective pretty strongly.  That's the idea.

Edited by ccaissie, 23 November 2018 - 10:11 PM.

[Richard O'Neill]

Are you certain the crown element hasn't been reversed?

Edited by Richard O'Neill, 23 November 2018 - 11:05 PM.

[Mark Harry]

If it was OVERcorrected, the spacing could be enlarged, and you could make a significant change. Sounds like a drastic action, modifying a surface is in order.
Too bad.
I would suggest -NOT- to mess with R3, nearest the crown---only consider R4 to be the one to get tweaked.
R2 and R3 have a very close tolerance (critical) in respect to each other. Likely varies by a set number of interference fringes.
If you enlarge R1, that could add some correction, but you'd add to the focal length fairly quickly as well.
*********
  That leaves your option you found--- by changing the conic of R3 (if it isn't coated)
If you miss, and go a bit too far, then you could use the spacing idea to tweak the correction to where you want. So there would be an option that could be called a "fudge factor".

[ccaissie]

Original pencil mark on crown and flint are matching.  R1 is longer than R2.  R2 is slightly shorter than R3, but close enough to create Newton's rings under mono light.  No such interference can be seen if R1 is contacted R3. 

[Mark Harry]

Correct.

[ccaissie]

If it was OVERcorrected, the spacing could be enlarged, and you could make a significant change. Sounds like a drastic action, modifying a surface is in order.
Too bad.
I would suggest -NOT- to mess with R3, nearest the crown---only consider R4 to be the one to get tweaked.
R2 and R3 have a very close tolerance (critical) in respect to each other. Likely varies by a set number of interference fringes.
If you enlarge R1, that could add some correction, but you'd add to the focal length fairly quickly as well.
*********
  That leaves your option you found--- by changing the conic of R3 (if it isn't coated)
If you miss, and go a bit too far, then you could use the spacing idea to tweak the correction to where you want. So there would be an option that could be called a "fudge factor".

Yes, I ran out of spacing options which produced little effect.  

 

The reason I'd mess with R3 is because I'm familiar with working and testing concave surfaces.  In OLSO it shows that just a bit of aspherization, like going from +2 to -2 makes a huge difference in the total SA.  I can work in this range.  At least I think I'm assessing this right.

 

I found this thread helpful.

 

https://www.cloudyni...ractor-project/

 

I'm sure that by just tweaking a conic, there are other less than optimal effects induced, but at this point, the SA is quite bad, and getting that 5th order spherochromatism and other HSA issues handled is not the first order of the day.

 

Just sayin'.

 

Yes.  My mechanical design mentor said that you should always have a 'hole card' or 'plan b'.  Spacing changes can provide that fine tuning.

Edited by ccaissie, 24 November 2018 - 06:53 AM.

[DAVIDG]

 Foucault test the concave element on the flint. It should be a perfect sphere and it is  an easy test. If it is good then  the problem is most likely R1 or R2 since R4 being almost flat doesn't have much optical power. 

   There is also the possibility that all surfaces are well figured and one or more of the radius is off since that will cause spherical aberration as well. As you know  the radius on an objective usually need to be within 1% or better of design  radii or it just won't work.  If a radius is too far off then you can't fix it be aspherizing a surface. You have expensive tool glass. 

  Also remember since a lens is a  refractive system that the errors on the surface are  the opposite of what is on a mirror. So for example a hill will look like hole.  A hill will cause the light to converge and come to focus short on a lens so it's  Foucault pattern is that of a hole on  a mirror. The opposite is true of hole on the lens which cause the light to diverge and focus long so it will look like a hill on a mirror.  So what looks like an  oblate figure actually  has the same basic shape as parabola or hyperbola on the surface(s) of the lens.

 

                     

                             Happy Holidays

                            - Dave  

Edited by DAVIDG, 24 November 2018 - 11:00 AM.

[ccaissie]

Reversing the crown, reversing the entire assembly, adjusting shims...all remains undercorrected.  Reversing the crown makes the undercorrection most severe....like 3/4" LSA.

 

By making and testing mirrors, I have a sense of how much SA is showing up.  I'd say 1 wave.  In OSLO, assuming standard Fraunhofer scrip, by aspherizing  R3, about like an f/3 paraboloid makes a big difference.

 

I find that a conic of -1 ...a paraboloid on the concave R3 surface does exactly what I would expect...causes spherical overcorrection.  Visualizing what a lens zonal error vs. a mirror is tricky...pencil and paper for crude raytrace helps. 

 

The question is "do I want to do this?"  
Owner said he'd trade me an 8" flat to check out the lens...

[Mark Harry]

Dave, you're a bit off.
R4 "can" have a large effect, since the flint element's power AS A WHOLE is dependent on the radii difference between both surfaces. (R3 & R4)
*****
Second, if I see a conic cropping up, then I -KNOW- for sure that can be solved with either radius change somewhere, or a spacing change. R3 is a common place for this to occur.
*********

Colin, I think you can mostly retain everything with this achro if you explore whether the conic can be adjusted on R3. But if you want to make a reasonable effort, I hope you use  a digital indicator, and "rings" or "cups" to make very accurate measurements of the radii you now have---as well as knowing the correct substrates used.
(???)
This would certainly improve your chances of being successful.

[davidc135]

Is the lens going back to the owner?  David

[ed_turco]

Professional optical designers always assume all spherical surfaces for any objective,.  In amateur hands, if there were slight differences from spheres on radii 1, 2 or 3, some fudging on R4 can bring light to its proper focus.

 

When you think about it, why would anyone want to design an objective with all nonspherical surfaces?

 

I think that one needs to rethink this Unitron objective's problems before sending it back.

[ccaissie]

I doubt that the design calls for aspherical surfaces.  Just that sloppy manufacture could have allowed a polished but uncorrected surface into the assembly.

 

Right that the other radii could be spherical but out of spec.  In OSLO, changing the value of R2 or R3 really changes the SA.

 

Snowing today, might have time to foucault test R3.  If it's oblate, that will be a significant clue to the undercorrection I see.

 

In a well corrected, all spherical Fraunhofer, by aspherizing R3 to just -0.1 overcorrects it dramatically

 

UPDATE:  R3 is a nice sphere. 

Edited by ccaissie, 27 November 2018 - 11:04 AM.

[ccaissie]

Settling into my shop now...most outside chores done.

 

I'm contacting the scope owner to see if he's game to have me aspherize R3 to correct this objective.  By working a conic into the surface, I can test it progressively and stop at a good point.

 

What are the guidelines for working a concave flint surface.  I plan to use a full diameter starred lap with straight burgundy pitch on the soft side.  Cerium?  Red rouge?  Workpiece on top, 1/3 blending stroke.

 

From the star test and DPAC I see undercorrection, stronger departure around the edges...the Ronchi lines inside focus are definitely pincushion curved, but they are much more strongly curved towards the edge.  I'll do zonal measurements to get a baseline.  Maybe photos...might even rig a camera stand for a change...

[DAVIDG]

 Just be sure you understand what the actual shape you need on the glass since this a refractive element. When OSLO shows a negative conic, the shape  is calculated with light being refracted through the element so the curve your polishing in can be  the opposite in shape of that of a mirror.

   From example when I make my Schupmanns I need to put a hyperbola on the objective to remove the undercorrection  but I'm not deepening the center like on a mirror but using a ring and inverted star lap to wear down the outer zone and raise the center. 

   If I was doing the work and I would test and correct R3 so it was a perfect sphere Now you know that surface isn't the problem. Test to see what correction is like and then if still under corrected put the correction on R1. 

 

                 Good Luck and Happy Holidays

 

                      - Dave 

[dan chaffee]

Based on what you've said, the lens has an undercorrected outer zone, or a rolled edge. If you are going to work R3, you need to turn the outer zones up and leave the inner zones alone. Flint corrects slowly....and scratches if you sneeze;-)

[ccaissie]

Yes undercorrected edge, it focuses short, the center focuses long.  The DPAC Foucault readings for the center, the 70% and the edge are 0, -.097, -.177.  The edge rays focus .177" closer to the lens than the center rays.

 

Since I Foucault tested R3... the concave surface and it is very spherical, the issue is a wrong radius or a badly figured surface.

 

In the Ronchi pics, I find the edge is indicated to be more strongly curved...we can say that one of the convex surfaces has a rolled edge, which increases the power there and makes it focus short.  I also notice that the very center focuses quite long...must be a "hole"  on one of the convex surfaces.

 

Pic on the left, pincushion is INSIDE.   The barrel shape is OUTSIDE.  100lpi

 

Undercorrected.

Attached Thumbnails

• 
• 

Edited by ccaissie, 14 December 2018 - 05:54 PM.

[ccaissie]

 Just be sure you understand what the actual shape you need on the glass since this a refractive element. When OSLO shows a negative conic, the shape  is calculated with light being refracted through the element so the curve your polishing in can be  the opposite in shape of that of a mirror.

   From example when I make my Schupmanns I need to put a hyperbola on the objective to remove the undercorrection  but I'm not deepening the center like on a mirror but using a ring and inverted star lap to wear down the outer zone and raise the center. 

   If I was doing the work and I would test and correct R3 so it was a perfect sphere Now you know that surface isn't the problem. Test to see what correction is like and then if still under corrected put the correction on R1. 

 

                 Good Luck and Happy Holidays

 

                      - Dave 

Yeah, I'm going over the test images and making experiments in OSLO.  When I put a -1 conic on R3 it overcorrects the assembly.  That's the direction I want to go in. 

 

But is a conic of -1 for  lens the same as a concave reflective paraboloid?   

 

NO.  surprise...thanks for your gentle queries helping me to finally get the 'reversing' of the sense of the lens power vs mirror.

My eagerness to parabolize R3 is misguided....It needs a turned up edge and a flatter center.

 

My eagerness to alter R3 is because I've never worked a convex surface...not familiar with testing a convex surface....Gaviola?

Edited by ccaissie, 14 December 2018 - 06:17 PM.

[ccaissie]

 Just be sure you understand what the actual shape you need on the glass since this a refractive element. When OSLO shows a negative conic, the shape  is calculated with light being refracted through the element so the curve your polishing in can be  the opposite in shape of that of a mirror.

   From example when I make my Schupmanns I need to put a hyperbola on the objective to remove the undercorrection  but I'm not deepening the center like on a mirror but using a ring and inverted star lap to wear down the outer zone and raise the center. 

   If I was doing the work and I would test and correct R3 so it was a perfect sphere Now you know that surface isn't the problem. Test to see what correction is like and then if still under corrected put the correction on R1. 

 

                 Good Luck and Happy Holidays

 

                      - Dave 

R3 is already a perfect sphere...  It looks like if I can make an oblate spheroid shape on R3, that will shorten the center rays and lengthen the edge rays....without changing the chromatic correction, which seemed ok.   A ring tool to deepen at the 80% zone? 

Edited by ccaissie, 14 December 2018 - 08:26 PM.

[Mark Harry]

Put in less than -1.0 for a conic, say -0.50. See where that gets ya.
That would be a prolate ellipse, and could be done rather quickly.

If you're using a spreadsheet of sorts, just multiply the paraboloidal ideal values by .5, if the correction would be correct for your situation.

[DAVIDG]

 If R3 is a  perfect sphere I wouldn't touch it. You have a very similar figure to what my Schupmann objectives start out with before I aspherize a surface.  What I would do is put the correction on R1. You need to deepen the area around the 70% zone and turn the edge up. The curve you want looks like what is put on a Schmidt corrector. So I would start out with a ring lap about 3/4" wide and centered over the 70% zone of. R1  Polish for a while and  you should see the Ronchi line straightening but the edge will get worse. Don't panic. Once you get the lines straight over center part of the lens, you switch a small star lap about the size of  a quarter. You polish with the center of the star  over the crest of the turned edge using a  circular stoke. The edge will  starts to reduce and move outward.  As it reduces you trim the tips of the star off and move the lap closer to the edge. You keep doing this until  your left with a small circular lap about the size of penny. Your most likely be left with tiny edge that is covered by the cell. 

   Here is a picture of a 6" objective that I figured using the above technique.  The lens was figured in the red. 

 

            - Dave 

Edited by DAVIDG, 15 December 2018 - 01:39 PM.

[ccaissie]

David:

 

Never have worked a convex surface.  A bit reluctant to do the assembly/ disassembly routine a whole lot, but ok.   I would think the stroke should go out almost to the edge...for a 4" lens, a ring about 3/4" wide and with a hole in the center about 1"?  So a pretty short stroke...like 1/4 d.    I plan on using soft pitch since I'm initially not dragging the lap over the edge.  

 

So how long would you work the surface with the ring lap using cerium?  10 minutes.... reassemble and test?

 

....Interesting that you use the term hyperbola for a lens .... not the actual shape of the surface, but the effect it has on the light rays....is that correct?

Edited by ccaissie, 15 December 2018 - 03:42 PM.

[DAVIDG]

 You have a defect surface(s) in the system. Right now you know the R3 is a sphere. So  when it comes to the total figure that one is not the problem ( that radius could be wrong) so the odds are it  is one of the other(s). So in my opinion your not  losing anything by working on one of the convex surfaces. 

   For the ring lap what I do is pour a full size lap and then make a mask that I press into it.  So the mask would be 4'" in  diameter with the ring section cut out into 3 or 4 sections with three or four spokes so  it is  all one piece. So when it is pressed into the pitch the open area will get raised up. You can then repress the lap back  into any shape if you need a need shape. I usually polish for about 5 minutes then test so things don't get away from me. Because you  have 4 surface involved it would be hard to just try to test one vs the whole lens to see if your going in the right direction. So I don't think your going to be able to get around not testing the complete lens after each polishing spell. 

  Hyperbola is the correct shape on the surface. It is just that you need to take that shape and invert it when it is on a convex surface vs a concave one.  Take a parabola for example, on a mirror you take material out of the center to deepen the sphere into a parabola. On the convex surface  you can't take material out of the center, you would have to add material in the middle to have same effect. Since you can't do that, you need to remove material from the  outer  section of the lens. So if you picture in this way, if  you start with a  sphere, and you want to turn that surface into asphere  on a concave surface you remove material so the asphere is on the outside of the starting sphere. On a convex surface you remove material so the asphere is on the inside of the starting sphere.  Hope this makes things a bit clear of how you visualize how you figure a convex vs a concave surface.  Good luck.

 

                        - Dave 

 

Edited by DAVIDG, 15 December 2018 - 06:31 PM.

[davidc135]

'A bit reluctant to do the assembly/disassembly a whole lot, but OK'. I'd make a simple wooden support that ensures the objective is square on to the flat, ready for testing.  David

[ccaissie]

'A bit reluctant to do the assembly/disassembly a whole lot, but OK'. I'd make a simple wooden support that ensures the objective is square on to the flat, ready for testing.  David

I'm not whining about the lens setup on the flat..only the repeated assembly of the lens in cell.  Each time it gets worked on is a chance of boo-boo.  Maybe I'll become an expert.

[DAVIDG]

 Here is  a picture of my commercial lens fixture that I use when double pass testing  my lens. You could make something very similar to hold the elements when you test your objective. I have taped the elements together when testing so I can quickly get them tested and  then get back to figuring. Good Luck.

 

                       - Dave